Osteosarcoma is the most common primary bone malignancy of childhood and adolescence and despite aggressive cytoctoxic chemotherapy and surgery, fully 40% of these children die of their disease. The cells comprising an osteosarcoma, and many other cancers, are functionally heterogeneous with regard to proliferation rate, tumorigenicity and metastatic ability. The biological basis for this heterogeneity remains unclear but has significant potential therapeutic implications, in that development of therapies directed at the biology supporting the more malignant cells might improve survival. Working to understand intra-tumoral heterogeneity, we have recently identified a highly tumorigenic osteosarcoma initiating cell visualized by its ability to activate an Oct-4 promoter/GFP reporter construct. This Oct-4/GFP+ population produces heterogeneous tumors comprised of both GFP+ and GFP- cells. Tumorigenic potential however, is maintained only in the GFP+ fraction, and is lost in the GFP- fraction, demonstrating a reversal of the malignant phenotype in the GFP- cells during tumor growth suggestive of epigenetic re-patterning. To identify the biological machinery responsible for driving the tumorigenic phenotype we performed differential analysis of the global expression patterns of the GFP+ and GFP- cells in tumors generated from a single GFP+ clone. The data were strikingly informative in that only 168 known genes showed significantly elevated transcription in the GFP+ cells, and of these at least 113 were directly linked to G2/M phase and spindle assembly checkpoints in the cell cycle. In contrast, 81 genes showed enhanced expression in the non-tumorigenic GFP- cells. The coordinate activities of these genes focused specifically at the endoplasmic reticulum stress response. The differences in the global expression patterns between the GFP+ and GFP- cell populations suggest two distinct molecular mechanisms supporting the disparity in their tumorigenic activity. Enhanced expression of G2/M and SA checkpoint genes enables the tumor-initiating cell to override these steps to proliferate at an accelerated rate, and conversely that cellular stress induces broad changes in the global transcriptional patterns of the GFP- cells, away from proliferation and toward survival pathways, reversing the malignant phenotype. For the present proposal, exploiting our ability to visualize the osteosarcoma initiating cell, we will work to determine the functional relationships between 1) the activation of stress pathways and malignant reversion, and 2) G2M/SA override and tumorigenicity. With the long term goal of improving survival for children with osteosarcoma we will address the following specific aims: AIM 1: To Determine if Environmental Stress Induces Malignant Reversion in the Osteosarcoma-Initiating Cells. AIM 2: To Determine if G2/M-SA Checkpoint Override is the Mechanism Supporting the Tumorigenic Activity in the Oct-4/GFP+ Cells. AIM 3: To Determine If G2/M-SA Checkpoint Control and ER Stress are Predominant Regulatory Pathways in the Tumorigenesis of Osteosarcoma.